June 1st, 2015, 1:49 pm
Update:I tried the 4d normal distribution with corr matrix == I. Even with NX = NY = NZ = NP = 20 I get initial results like 1, 0.5, 0.25, 0.125, 0.0625.This certainly proves feasibility of Goursat. As before, the spectrum of values is in a 4d tensor for lookup and interpolation. I took the code from the 3d case and extended it with one extra dimension. it took ~ 30 minutes to update the code.
Last edited by
Cuchulainn on May 31st, 2015, 10:00 pm, edited 1 time in total.